{
  "slug": "command-a-218b-a25b",
  "order": 73,
  "name": "Command A+ (218B-A25B)",
  "baseTitle": "Command A+ 218B-A25B",
  "company": "Cohere",
  "country": "Canada",
  "flag": "🇨🇦",
  "date": "2026-05-20",
  "decoderType": "Sparse MoE",
  "attention": "16:1 GQA with 3:1 sliding-window/global attention",
  "layerMix": "24 sliding-window + 8 global GQA/MoE",
  "scale": "218B total, 25B active/token (11.5% active)",
  "totalParams": 218000000000,
  "activeParams": 25000000000,
  "context": "128K input, 64K output",
  "contextTokens": 131072,
  "license": "Apache License 2.0",
  "kvCacheText": "128 KiB",
  "kvCacheKiB": 128,
  "kvCacheLevel": "Moderate",
  "summary": "Cohere's open sparse MoE model scales the Command A parallel transformer block to 218B total parameters while keeping 25B active per token.",
  "keyDetail": "Runs attention and MoE in parallel, which Cohere reports gives similar quality to a sequential block with better throughput.",
  "image": {
    "full": "https://sebastianraschka.com/llm-architecture-gallery/images/architectures/command-a-plus.webp",
    "thumb": "https://sebastianraschka.com/llm-architecture-gallery/images/architectures/thumbnails/command-a-plus.webp",
    "alt": "Figure 1: Command A+ 218B-A25B",
    "width": 960,
    "height": 939
  },
  "articleUrl": null,
  "links": [
    {
      "label": "config.json",
      "href": "https://huggingface.co/CohereLabs/command-a-plus-05-2026-w4a4/blob/main/config.json",
      "title": "CohereLabs/command-a-plus-05-2026-w4a4"
    },
    {
      "label": "Tech report",
      "href": "https://cohere.com/blog/command-a-plus",
      "title": null
    }
  ],
  "concepts": [
    {
      "slug": "gqa",
      "label": "GQA",
      "name": "Grouped-Query Attention",
      "url": "https://llm-atlas-7or.pages.dev/concepts/gqa/"
    },
    {
      "slug": "moe",
      "label": "MoE",
      "name": "Mixture of Experts",
      "url": "https://llm-atlas-7or.pages.dev/concepts/moe/"
    },
    {
      "slug": "nope",
      "label": "NoPE",
      "name": "No Positional Embeddings",
      "url": "https://llm-atlas-7or.pages.dev/concepts/nope/"
    }
  ],
  "aai": {
    "v5": {
      "label": "AA Intelligence Index v5.0",
      "total": 6.2,
      "profile": {
        "Knowledge Work": 0.6,
        "Coding": 0.2,
        "Science": 1,
        "General": 23.2
      },
      "url": "https://artificialanalysis.ai/models/command-a-plus"
    },
    "v4": {
      "label": "AA Intelligence Index v4.3.2",
      "total": 13.1,
      "profile": {
        "General": 26.4,
        "Scientific": 6.1,
        "Coding": 19.5,
        "Agents": 0.5
      },
      "url": "https://artificialanalysis.ai/models/command-a-plus"
    }
  },
  "sourceUrl": "https://sebastianraschka.com/llm-architecture-gallery/#card-command-a-218b-a25b",
  "family": "moe",
  "url": "https://llm-atlas-7or.pages.dev/models/command-a-218b-a25b/",
  "markdownUrl": "https://llm-atlas-7or.pages.dev/models/command-a-218b-a25b.md"
}